This Bull Itanium2 system provides 192 Itanium2 processor cores available for researchers on campus. The system is configured as 24x compute nodes, each with 4x Intel Itanium2 Montecito Dual Core 1.6GHz/8MB cache (i.e. 8 cores per node) and 16GB RAM. Each core has a peak compute performance of 6.4 GFlop/s, giving a total peak performance of about 1.2 TFlop/s. Each node also has up to 512GB of local scratch disk. Four of these nodes are configured as high memory nodes and have 32GB RAM. All of the nodes are connected by a single rail Quadrics QsNetII (elan4) interconnect (900 Mbytes/second sustained bandwidth). The system has up to 10TB of central filestore based on Lustre (a parallel filesystem).
MPI is supported across the whole system and OpenMP within the 8 core nodes. A variety of commonly used codes are installed on the system, of course if you have a requirement for a code please contact us. Performance analysis and debugging tools specifically for HPC systems will be available. RCS provides training courses on all aspects of using Horace efficiently.
While the system will operate predominantly as a batch service, interactive use will be available for code development, and special arrangements can be made for novel use of the service such as computational steering.
Tel: +44 (0)161 275 6408